Four stories, one move

If you want to know how I work, method names won’t get you far. Cases will. Here are four: what the brief was, what everyone treated as settled, and what happened once the settled part turned out to move.

The same move runs through all four stories. I name it at the end rather than up front — otherwise you stop reading the cases and start reading evidence.

AOK: the consent that wasn’t a pop-up

2016/17, retargeting for a German health insurer. Two fears were on the table from the first day. The first: how can a German health provider work with a US marketing company and stay on the right side of the law? The second came from experience and weighed more: every complex project before this one had died on the last mile, at the data protection officers.

Two sentences counted as immovable. The first: privacy reviews at the end. We had all 16 regional data protection officers in the room from day one, rather than handing them a finished concept to sign off. The body that says no at the end became people who had helped build it.

The second: consent is a pop-up that opens before the first page does. That pop-up manufactures the resistance it then measures. We put the consent into a questionnaire form instead — an ordinary checkbox, inside a form only the actual target audience ever filled in. Acceptance rose significantly.

Miele: more technology does not fix a selling problem

The brief arrived as a question: what could we implement across different shop systems in different countries to raise the conversion rate? It sat below one percent.

I turned the question around. Why are e-commerce systems so technically driven in the first place? They have more tech DNA than sales DNA. A shop can run flawlessly and still not sell: presenting products and processing orders is not the same thing as selling. Offline that craft was taken for granted; online it appeared nowhere in the flow.

So I brought in a salesperson from the offline business and translated their craft into the online flow: needs analysis, objection handling, the classic principles of selling, only in pages and steps instead of in conversation.

Digital revenueRevenue is the money that comes in before anything gets subtracted — 100 ice creams at €1 each is €100 in revenue. It tells you how much business happened, not whether the day was worth it. full card went up tenfold. Not because the technology got better, but because somebody was finally selling.

Payback: the right question at the wrong moment

Payback is an offline system at heart, and its trigger is one sentence at the till: “Do you collect points?” It works because the shopping is already on the belt.

For the digital world it was taken as given that the question belongs where it works offline: at the checkout. But online, whether and where someone buys is settled earlier. Ask at the digital till and you are asking after the decision.

So we moved the question ahead of the visit. A toolbar intercepted navigation and told people they could earn points on the site they were on their way to. Baskets got bigger and the conversion rate went up.

The concept was sold. What came out of it is the Payback app that exists today.

FriendScout24: from 600 a day to 18,000

Marketing was running weekly sprints: test, read the numbers, next variant. The rhythm broke on a piece of arithmetic. A banner from the agency cost over €400 — per variant. Test weekly and that is a budget nobody signs off.

The debate was about test frequency. The bottleneck was not the frequency, it was the price of a single variant. So we hired a graphic designer in-house instead of buying creative by the piece. After that a variant cost hours instead of hundreds of euros, and the weekly rhythm held.

About six months and roughly 25 iterations later we knew from the data which mechanics actually drove click-through and conversion. Registrations went from 600 a day to 18,000, thirty times over. What found it was not the best idea. It was the twenty-fifth round.

The move

Four projects, four industries, four different briefs. The leverage sat in the same place every time: not in the execution, but in a sentence nobody had touched. Privacy reviews at the end. Consent is a pop-up. E-commerce is a technology question. Loyalty starts at the till. Creative comes from the agency.

None of those sentences was wrong. They were simply not as fixed as everyone assumed. And because they counted as fixed, the optimizing happened beside them — on the layout, the targeting, the test frequency. Everywhere movement was permitted and there was little left to win.

If you want to know what I actually do inside a project, it is this one sentence:

I move the assumption everyone treats as fixed.